The role of the gold standard in Keynesian monetary theory.
The primary motivation for this paper was provided by ambiguities within the General Theory regarding the efficacy of monetary policy remedies to a depression. I will argue that rather than being a "fixed-price" or "depression" model, the General Theory is best thought of as a "gold standard" model. Thus, I show that Keynes's statements on monetary policy ineffectiveness implicitly assumed the existence of a monetary system with some linkage to gold and that some of the General Theory's most distinctive features make more sense if viewed from that perspective.
This hypothesis may appear surprising in light of the popular perception that in the General Theory Keynes tried to move away from the open-economy, international gold standard framework employed in the Treatise on Money. I will argue that he failed to do this. Even if Keynes did not explicitly regard the General Theory as a gold standard model, during the 1930s he began to view the economic problems that we associate with the gold standard (monetary policy ineffectiveness, occasional episodes of deflation, and so on) as being inherently characteristic of a modern capitalist economy. Ironically, Keynes made this change just as many countries were beginning to deemphasize the role of gold.
Keynes is regarded as one of the most severe critics of an international gold standard with rigid parities. Thus, it is odd that he would develop a monetary ineffectiveness proposition that has little or no applicability to a country operating under a fiat money regime. An essential part of my argument will be to show that Keynes viewed a pure fiat regime as being both highly unlikely and highly undesirable.
An important implication of this paper is that the gradual replacement of the gold standard by fiat money regimes should have led modern Keynesian economists to move away from some of the most distinctive features of 1930s Keynesianism (such as monetary policy ineffectiveness, deficit spending, the paradox of thrift, and protectionism.) I will show that the recent metamorphosis of New Keynesian macroeconomics represents a natural response of economists to a radically altered monetary environment.
II. KEYNES'S VIEWS ON MONETARY POLICY
To better understand the role played by the gold standard in Keynes's work, it will be helpful to start with the "monetary ineffectiveness proposition." In current usage, this phrase generally refers to the New Classical view that money has no impact on real output. In contrast, most would interpret the pessimistic views expressed by Keynes during the 1930s as reflecting a belief that monetary expansion might be incapable of increasing nominal output. Because Keynes clearly did not believe that an expansionary monetary policy would reduce real output, this interpretation suggests that Keynes viewed the ineffectiveness proposition as also applying to the price level. Unfortunately, there is great uncertainty as to how pessimistic Keynes actually was regarding the effectiveness of monetary policy.
The 31 December 1933 issue of the New York Times [Times, p. 2XX] published a long letter by Keynes in which he evaluated the first year of the New Deal. After a lengthy critique of the National Recovery Act, Keynes turned his attention to Roosevelt's gold-buying program:(1) "The other set of fallacies . . . arises out of a crude economic doctrine commonly known as the quantity theory of money. Rising output and rising incomes will suffer a setback sooner or later if the quantity of money is rigidly fixed. Some people seem to infer from this that output and income can be raised by increasing the quantity of money. But this is like trying to get fat by buying a larger belt." This is a clear example of what is often referred to as the "pushing on a string" view of monetary policy ineffectiveness. But notice how Keynes subtly shifts his argument in the very next paragraph, as he turns from money's effect on real output to its effect on prices: "It is an even more foolish application of the same ideas to believe that there is a mathematical relation between the price of gold and the price of other things. It is true that the value of the dollar in terms of foreign currencies will affect the prices of those goods which enter into international trade. In so far as an overvaluation of the dollar was impeding the freedom of domestic price-raising policies or disturbing the balance of payments with foreign countries, it was advisable to depreciate it. But exchange depreciation should follow the success of your domestic price raising policy as its natural consequence, and should not be allowed to disturb the whole world by preceding its justification at an entirely arbitrary pace." Here Keynes switches from a positive to a normative argument. And he carefully avoids applying the monetary ineffectiveness proposition to prices. Because the U.S. price level did rise rapidly during 1933, it is not surprising that Keynes would careful qualify his assertions regarding prices.
Keynes also indicated that he remained a proponent of a managed currency but then stated that "the recent gyrations of the dollar have looked to me more like a gold standard on the booze than the ideal managed currency of my dreams." At times he sounded much like Roosevelt's conservative critics: "This game of blind man's bluff with exchange speculators serves no useful purpose and is extremely undignified. It upsets confidence, hinders business decisions, occupies the public attention in a measure far exceeding its real importance, and is responsible both for the irritation and for a certain lack of respect which exist abroad."
The preceding statements raise several important questions. Did Keynes oppose the adoption of a highly expansionary monetary policy during 1933, or was he merely opposed to the specific policy tools adopted by Roosevelt? Did Keynes really believe that an expansionary monetary policy would not be effective in promoting recovery? Did Keynes regard Roosevelt's policy as being inflationary, and, if so, was that viewed as being a desirable attribute?
A rough draft of the General Theory was completed by the spring of 1934, and thus it is not surprising that some of the ideas (and ambiguities) appearing in Keynes's Times letter also appear in the General Theory. One way of summarizing the preceding ambiguities is to note that in the General Theory Keynes appeared to have simultaneously held the following three views:
1. During a depression, monetary policy may be unable to restore full employment.
2. Sustained inflation will only occur if aggregate demand increases sufficiently to move the economy to a position near, or at, full employment.
3. Under a fiat money system, it is always possible to generate sustained inflation through a suitable expansion of the money supply (and/or currency depreciation).
The preceding assertions are logically inconsistent. If it is always possible to generate sustained inflation with a sufficiently expansionary monetary policy, and if sustained inflation only occurs as the economy approaches full employment, then why would Keynes argue that an expansionary monetary policy might be unable to promote recovery from a deep depression? Did Keynes, then, actually hold all three of the views expressed above?
The first two assertions represent two of the most important elements of the General Theory.(2) The third statement is much more problematical. In Keynes's writings from the 1930s one can find evidence both for and against the view that he intended the policy ineffectiveness proposition to apply to prices. I will attempt to show that what appear to be logical contradictions in Keynes's views on monetary policy effectiveness are probably the result of differing assumptions as to the constraining influence of the international gold standard.
III. MONETARY POLICY INEFFECTIVENESS: PRICES OR OUTPUT?
One of the unusual features of the General Theory is that discussion of the impotence of monetary policy always involves output rather than prices. Perhaps Keynes believed in both forms of policy ineffectiveness but found no need to note specifically the implication of policy ineffectiveness for changes in the price level. There are reasons, however, to doubt this interpretation.
Throughout his career, and especially in the General Theory, Keynes relished taking positions at variance with the prevailing orthodoxy. And during the interwar period much of orthodox opinion, both inside and outside of academia, was skeptical of the view that an expansionary monetary policy could cure a depression. At the same time, orthodox opinion also opposed monetary expansion on the grounds that it would be inflationary. If Keynes wanted to sharply delineate the differences between his views on monetary policy and the prevailing orthodoxy, he would have been expected to emphasize the hypothesis that the adoption of a highly expansionary monetary policy during the 1930s would be ineffective in raising prices, not output.
Meltzer  observed that, throughout his career, Keynes had been a consistent foe of inflation. Keynes had observed and written about the various post-World War I inflations: "Lenin is said to have declared that the best way to destroy the capitalist system was to debauch the currency. . . . In the latter stages of the war all the belligerent governments practised, from necessity or incompetence, what a Bolshevist might have done from design. Even now, when the war is over, most of them continue out of weakness the same malpractices" [1982 (1919), Collected Writings, vol. 9, 57]. To some extent this remark reflects the prevailing issues facing European policymakers in 1919. There is evidence, however, that even in the 1930s Keynes continued to have a strong aversion to policies that would "debauch the currency." In the Treatise [1953 (1930), vol. I, 170], Keynes called an unmanaged gold standard "the worst of all conceivable systems (apart from the abuses of a fiat money which has lost all its anchors)." And in the midst of the Great Depression, Keynes congratulated Roosevelt for refixing the dollar to gold and thus not succumbing to the views of what he called the "extreme inflationists" [1982 (1919), Collected Writings, vol. 21, 312]. Whether one regards these remarks as being reflective of Keynes's policy preferences, the rhetoric is certainly not what one would expect from someone who denied the ability of governments to generate inflation, if they so desired.
In the General Theory  Keynes comes close to conceding that the monetary policy ineffectiveness proposition only applies to real output: "So long as a tolerable level of employment could be attained on the average of one or two or three decades merely by assuring an adequate supply of money in terms of wage-units, even the nineteenth century could find a way. If this was our only problem now - if a sufficient degree of devaluation is all we need - we, to-day, would certainly find a way."(3) Note the similarity to the Times quote at the beginning of this section. Once again, Keynes seems to apply the policy ineffectiveness proposition to output but not to prices.(4)
In the previous discussion I have argued that Keynes seemed reluctant to extend the policy ineffectiveness proposition to prices but that an implication of his theoretical work is that it should be equally applicable to prices and output. Did Keynes fail to understand that implication? The following quotation from the Treatise suggests that Keynes did understand the logic of his theory: "Not until deliberate and vigorous action [i.e. expansionary open market operations] has been taken along such lines as these and has failed, need we, in the light of the argument in this Treatise, admit that the Banking System can not, on this occasion, control the rate of investment and, therefore, the level of prices" [Treatise, vol. II, 387]. Here Keynes does extend the monetary ineffectiveness proposition to prices, at least on theoretical grounds.
I know of no satisfactory way of reconciling all of Keynes's various remarks regarding policy ineffectiveness and prices. What I hope to show is that Keynes's apparently inconsistent views on monetary policy effectiveness may reflect differing assumptions regarding the extent to which policy was, or should be, constrained by a link to the international gold standard. What makes this exercise so difficult is that Keynes himself often seemed oblivious to this connection. (His pessimism over monetary policy expressed in the preceding Times quotation occurred during a period when the United States had abandoned the gold standard.) I will show that some of the key concepts in the General Theory are based on a macroeconomic environment that modern economists would associate with the gold standard but that Keynes simply assumed were characteristic of a market economy. Some of the most important characteristics of the classical gold standard were the exogeniety of money and prices, the lack of persistent inflation, and the lack of inflation expectations. Thus, we need to look for evidence linking Keynes's views to that type of environment.
IV. THE ABSENCE OF INFLATION EXPECTATIONS IN KEYNES'S WORLDVIEW
Keynes's policy environment differed radically from that of modern economists brought up under fiat money regimes. I will attempt to show that many of Keynes's assertions about monetary policy can best be understood if we assume that Keynes never gave serious consideration to the adoption of monetary policies that would result in a significantly positive expected rate of inflation.
The preceding hypothesis may seem strange to a contemporary economist accustomed to persistent inflation. In the century prior to 1936, however, no nation had been able to generate sustained periods of the sort of persistent but moderate inflation associated with the postwar era, and thus Keynes had no firsthand observation of that sort of monetary policy. Keynes was very much a product of the gold standard environment.
Barsky  demonstrated that under the classical gold standard the price level was approximately a white noise process with zero trend and that even during periods when, ex post, a trend may appear discernible, the rational expectations-expected rate of inflation was near zero. Thus, for the purposes of this paper it will be assumed that a monetary system where the currency was expected to maintain a stable value vis-a-vis gold is equivalent to a monetary system with approximately zero expected inflation. (I will use the term "quasi-gold standard" to designate a policy expected to maintain a relatively stable nominal price of gold.)
Despite this historical pattern, one might imagine that Keynes's well-known opposition to the gold standard would have led him to advocate a policy of modest inflation. In fact, the following quotation shows that Keynes's opposition to the gold standard was far from absolute: "At all stages of the post-war developments the concrete proposals which I have brought forward from time to time have been based on the use of gold as an international standard, whilst discarding it as a rigid national standard. The qualifications which I have added to this have been always the same, though the precise details have varied; namely (1) that the parities between national standards and gold should not be rigid, (2) that there should be a wider margin than in the past between the gold points, and (3) that if possible some international control should be formed with a view to regulating the commodity value of gold within certain limits" [1982 (1933), Collected Writings, vol. 21, 186]. This is, of course, is quite similar to what was eventually implemented at Bretton Woods. Although the monetary system envisioned by Keynes would provide for somewhat greater policy flexibility than the classical gold standard, in 1933 Keynes had no reason to believe that the secular trend in prices would have been much different from the near 0% trend of the classical gold standard.
Keynes certainly understood the impact of highly expansionary monetary policies on expectations. In the Tract on Monetary Reform  Keynes had discussed the economics of hyperinflation, including the impact of expected inflation on money demand. But when Keynes discussed monetary policy options he either ignored options where these effects would be important or treated them as pathological cases unworthy of serious consideration. For instance, in his discussion of the quantity theory of money Keynes argues that, in general, the elasticity of prices with respect to changes in the money supply is less than one, and is equal to one in the (classical) full employment case. He then makes the somewhat odd assertion that "But in general e [that is, (dP/dM)/(P/M)] is not unity; and it is, perhaps, safe to make the generalization that on plausible assumptions relating to the real world, and excluding the case of a 'flight from the currency' . . . e is, as a rule, less than unity" [General Theory, 306]. It is interesting that Keynes regards this case as being so implausible that it can be safely ignored when considering plausible monetary policies. The phrase "flight from the currency" merely refers to an increase in inflation expectations leading to a drop in money demand. (A case that modern economists would view as being equally likely to occur as a disinflation-generated increase in money demand.) A few chapters later Keynes explicitly links this phrase with the extreme post-World War I inflationary episodes: "flight from the currency such as occurred during the post-war European currency collapses" [General Theory, 329].
The preceding views are in large measure a reflection of Keynes's skepticism regarding the practical importance of the Fisher effect. Prior to World War Il, the term "inflation" often referred to a once-and-for-all increase in the price level, and need not have been associated with a Fisher effect. But in discussing the Fisher effect with Robertson, Keynes even denies that expected inflation would have any impact on the nominal interest rate: "Now, as against this, my contention is that the expectation of a change in price has no effect on the rate of interest, since it leaves unchanged the relative attractions of cash and loans. Are you denying this? Do you think that an expectation of higher prices in [the] future causes the rate of interest to rise?" [1982 (1935), Collected Writings, vol. 13, 518].
Meltzer [1988, 128, footnote] is one of the few to recognize the important role played by (the lack of) inflationary expectations in Keynes's views on monetary policy ineffectiveness. He noted that a policy of inflation would reduce the real demand for money and move the economy out of a liquidity trap. He also found it strange that neither Keynes nor his critics had seemed to recognize this point. (Here Meltzer implicitly accepts the interpretation that Keynes's view of monetary policy ineffectiveness did not apply to prices.)
Meltzer's observation is certainly interesting but it does not really provide a solution to the liquidity trap. Whereas a modern economist might view a policy of inflation as a rather innocuous assumption, there is a sense in which by assuming a positive rate of inflation, one is assuming away not just the liquidity trap, but the whole concept of depression as Keynes understood the term. In the General Theory, Keynes viewed inflation as a symptom of an economy approaching full employment and suffering from "bottle-necks." Thus, Keynes would have been confused by Meltzer's suggestion that inflation was a way out of the liquidity trap. Keynes viewed an exit from the liquidity trap as a precondition for a recovery of output and recovery of output as a precondition for inflation.
To summarize, we know that during the 1930s Keynes became very pessimistic about the prospects for using monetary expansion to promote recovery and that on theoretical grounds this pessimism should have extended to prices as well. At the same time, Keynes was very aware of the inflationary potential of an unrestrained fiat regime. Perhaps Keynes ignored these regimes in his statements about monetary ineffectiveness because he thought a fiat regime to be both an unlikely and an undesirable policy choice. This would also explain his blind spot toward the problem of inflation expectations.
To develop the preceding into something more than a tentative hypothesis, however, we need to examine Keynes's views of contemporaneous macroeconomic issues. Fortunately, Keynes's statements regarding U.S. monetary policy during the early 1930s provide strong support for two of the essential arguments in this paper; that Keynes opposed inflationary fiat regimes and that Keynes's views on monetary policy ineffectiveness were implicitly linked to the constraints imposed by the international gold standard.
V. KEYNES'S VIEWS ON U.S. MONETARY POLICY, 1932-34
During depressed periods such as the early 1930s, Keynes would occasionally advocate a policy of "reflation," that is, raising prices back to their 1929 levels. But throughout his entire career Keynes was strongly opposed to the type of expansionary monetary policies that would lead to high or persistent inflation. At times he took positions that appear surprisingly conservative from a modern perspective. For instance, Keynes approved of Roosevelt's decision to raise definitively and then permanently fix the price of gold in January 1934: "He has adopted a middle course between old-fashioned orthodoxy and the extreme inflationists. I see nothing in his policy which need be disturbing to business confidence" [1982 (1934), Collected Writings, vol. 21, 312]. The "old-fashioned orthodoxy" were clearly the conservatives who opposed any devaluation. Keynes had called Roosevelt "magnificently right" when he torpedoed the World Economic Conference in July 1933 but had also argued that the U.S. should fix the price of gold at $28 an ounce. Roosevelt's decision to raise the price of gold to $35 an ounce represented a much sharper devaluation than Keynes had desired, but Keynes was relieved that the uncertainties associated with the gold-buying program would be brought to a definitive conclusion.
His comment about "extreme inflationists" is even more revealing and deserves further scrutiny. This term of approbation was widely used in the United States during the early 1930s and usually referred to the advocates of "greenbacks." In June 1933, Congress gave Roosevelt the authority to issue up to $3 billion dollars in greenbacks, but Roosevelt hoped to avoid this "radical" policy by using dollar depreciation to raise prices back to their pre-depression levels. Of course the $35 an ounce gold price was far too low to raise prices to the desired levels, and a $28 an ounce price would have been even more inadequate. It is not even clear that $3 billion in greenbacks would have been sufficient.(5) Why then would Keynes apparently view the issuance of greenbacks as extreme inflation? Did he view this as an example of "a fiat money which has lost all its anchors"?
In addition to greenbackism, there are at least two other ways of visualizing an expansionary monetary policy, reductions in the price of money (currency depreciation), and reductions in the rental cost of money (lower interest rates.) It is possible that the preceding comment on the "extreme inflationists" was a reference to the dollar depreciation program of late 1933. If so, then his opposition may have been partially motivated by the view that it represented a "beggar-thy-neighbor" policy of currency depreciation.
Throughout his career Keynes visualized the transmission mechanism between money and output primarily through changes in the nominal interest rate, rather than the exchange rate or the quantity of money: "If you are held back [that is, reluctant to buy bonds], I cannot but suspect that this may be partly due to the thought of so many people in New York being influenced, as it seems to me, by sheer intellectual error. The opinion seems to prevail that inflation is in its essential nature injurious to fixed income securities. If this means that an extreme inflation such as is not at all likely is more advantageous to equities than to fixed charges, that is of course true. But people seem to me to overlook the fundamental point that attempts to bring about recovery through monetary or quasi-monetary methods operate solely or almost solely through the rate of interest and that they do the trick, if they do it at all, by bringing the rate of interest down" [1982 (1934), Collected Writings, vol. 21, 31%20].
At first glance, the preceding passage may seem to contradict the view that Keynes opposed inflation. But in line 3 of the quotation Keynes seems to use the term "inflation" to represent a one-time reflation of prices. What is particularly striking about this passage is that Keynes suggests (in line 4) that only an "extreme inflation" would be sufficient to produce a Fisher effect. During November 1933, Roosevelt's gold purchase program was perceived by the financial press (but not necessarily by Keynes) as producing just such an effect and was opposed by orthodox opinion for that reason.(6)
The preceding quotation also contains many of the ambiguities discussed above. Keynes is clearly exempting the "extreme inflation" scenario from his blanket statement that expansionary monetary policies always reduce nominal interest rates. Here he implicitly recognizes the possibility of a Fisher effect. There remains, however, an ambiguity as to whether the "extreme inflation" exemption also applies to the hypothesis that an expansionary monetary policy can only promote recovery by reducing interest rates.
It should be noted that Keynes did, on numerous occasions, advocate a policy of open market purchases. But even in those cases he often seemed to go to extraordinary lengths to fashion a noninflationary open-market policy: "I see no reason why you should not reduce the rate of interest on your long-term government bonds to 2 1/2% or less; with favorable repercussions on the whole bond market, if only the Federal Reserve System would replace the present holdings of short-dated Treasury issues by purchasing long-dated issues in exchange" [Times, 31 December 1933, 2XX]. Of course, the effectiveness of such a policy is very doubtful.
The preceding quotations regarding Keynes's views on U.S. monetary policy contain one additional puzzle. In the December 1933 Times article Keynes argued that monetary policy cannot restore full employment, whereas less than three weeks later he is expressing relief that Roosevelt avoided the policy recommendations of the "extreme inflationists." One explanation for the preceding conundrum is that when applying the policy ineffectiveness proposition to output, Keynes was only considering a set of monetary policies that would be "reasonable," that is, unlikely to lead to a "flight from the currency" (or a positive expected rate of inflation). Conversely, when discussing inflation Keynes considered a wider range of policy options. I know of no instance where Keynes explicitly discusses the impact of an "extreme inflationist" policy on output, for a severely depressed economy.
The fact that Keynes opposed inflationary fiat regimes does not, by itself, prove that the General Theory was a gold standard model. The strongest support for that hypothesis comes not from his policy statements but rather from within the General Theory itself.
VI. MONETARY POLICY AND THE GOLD STANDARD
Many of Keynes's apparently contradictory statements about monetary policy can be at least partially reconciled if we assume that Keynes generally envisioned monetary policy operating within the institutional framework of a gold (or quasi-gold) standard. For instance, in his discussion of the role that expectations play in the liquidity trap Keynes specifically refers to an international gold standard setting: "But the long-term rate may be more recalcitrant when once it has fallen to a level which, on the basis of past experience and present expectations of future monetary policy, is consider 'unsafe' by representative opinion. For example, in a country linked to an international gold standard, a rate of interest lower than prevails elsewhere will be viewed with a justifiable lack of confidence" [General Theory, 203].
Keynes's skepticism about the existence of a complete liquidity trap is shown in one of the most frequently cited statements in the General Theory : "But whilst this limiting case might become practically important in the future, I know of no example of it hitherto." A less well known but more revealing quotation occurs just a few lines later where the only specific, real world, example of policy ineffectiveness provided in the entire General Theory refers to a country rigidly fixed to the international gold standard: "The most striking examples of a complete breakdown of stability in the rate of interest, due to the liquidity function flattening out in one direction or the other, have occurred in very abnormal circumstances. . . . in the United States at certain dates in 1932 there was a crisis of the opposite kind - a financial crisis or crisis of liquidation, when scarcely anyone could be induced to part with holdings of money on any reasonable terms" [207-8].
The term "certain dates" almost certainly refers to the spring of 1932, when an aggressive policy of open-market purchases by the Federal Reserve was associated with a run on the dollar, a massive outflow of gold, falling commodity prices, and a collapse in stock prices. Although this looks a lot like a liquidity trap, it is a very unusual variant of that phenomenon. The heavy gold outflows in the spring of 1932 prevented the Fed's open-market purchases from expanding the money supply. If the term "liquidity trap" refers to a situation where an increase in the money supply fails to lower interest rates or expand aggregate demand, then it is odd that Keynes would choose a period when the money supply did not expand as a way of illustrating that phenomenon.
Bernanke  used the phrase "multiple monetary equilibria" rather than "liquidity trap" to describe the situation faced by the United States during the early 1930s. Bernanke argued that under an international gold standard a loss of confidence in the future parity of currencies would lead to the hoarding of gold, a reduction of the money/gold ratio and thus a reduction in the money supply. One implication of this hypothesis is that expansionary monetary policy actions could lead to a reduction in confidence and a contraction in the money supply. Note that this problem would not occur under a fiat money regime. Thus, if episodes like the spring 1932 open-market purchases led to Keynes's pessimism about the effectiveness of monetary policy, then it would appear that this pessimism was intrinsically linked to the constraints imposed under an international gold standard regime.
The preceding interpretation can help illuminate a puzzling discussion of liquidity preference in the General Theory : "Circumstances can develop in which even a large increase in the quantity of money may exert a comparatively small influence on the rate of interest. For a large increase in the quantity of money may cause so much uncertainty about the future that liquidity-preferences due to the security-motive may be strengthened; whilst opinion about the future of the rate of interest may be so unanimous that a small change in present rates may cause a mass movement into cash. . . if we are to control the activity of the economic system by changing the quantity of money, it is important that opinions should differ. Thus the method of control is more precarious in the United States, where everyone tends to hold the same opinion at the same time." To a modern economist, this quotation must seem somewhat perplexing. When one thinks about large increases in the quantity of money creating uncertainty, "flights from a currency" and hyperinflation come to mind. But the references to currency hoarding, the failure to "control the level of economic activity" and to the situation in the United States, all suggest that Keynes had in mind a deflationary scenario such as was experienced by the United States during 1932. This scenario would require the existence of a monetary policy constrained by a fixed exchange rate regime.
Where does the preceding analysis leave Keynes's hypothesis of monetary policy ineffectiveness? Was he merely asserting that an expansionary monetary policy pursued under a gold standard regime might prove incapable of stimulating aggregate demand? Even Keynes's critics could accept that assertion without too much difficulty. Unfortunately, Keynes's views cannot be so easily reconciled. We have seen that Keynes strongly opposed the implementation of a fiat money regime. And we have also seen that Keynes's view of the liquidity trap was intimately linked to the experiences of countries constrained by a rigid gold standard. But there is one final problem that needs to be addressed: Keynes did favor a modified gold standard under which countries would be free to devalue their currencies as required to pursue a policy of reflation. Given his advocacy of this type of regime, and given the fact that the international gold standard began to collapse during the early 1930s, why would he have remained so pessimistic as to the prospects of using expansionary monetary policy to promote recovery?
In many respects, the international gold standard did not end in September 1931. During the mid-1930s major economies such as the United States, France, and Germany continued to peg their currencies to gold at a fixed parity. And after an initial depreciation during late 1931, the British pound maintained a relatively stable value against those currencies. Thus, after a one-time increase associated with devaluation, the world price level (and hence the expected rate of inflation) resumed a path not unlike what had occurred under the classical gold standard. Keynes still had little reason to surmise that even a modified gold standard could promote persistent inflation.
Although most countries continued to maintain noninflationary policies during the 1930s, there were important advocates of much more expansionary policies. On numerous occasions Keynes argued that he could support currency depreciation if it was supported by economic fundamentals: "If in the next two years President Roosevelt proves himself a more devoted, or at least a more successful, price-raiser than Mr. Neville Chamberlain, this will justify a still lower exchange value for the dollar. If they are both equally successful, the exchange value of the two currencies will remain, as it should, stable. . . . This arrangement would overcome the risk of unhealthy competitive exchange depreciation . . . America can fairly claim that, if she takes her price-raising programme more seriously than we do ours . . . she is entitled to an international value of the dollar which will not act as a drag on the success of her domestic policies" [1982 (1933), Collected Writings, vol. 21, 261-62]. Thus, there is a sense in which Keynes's believed that monetary policy during a depression should not be hamstrung by exchange rate constraints. But exactly how would such a policy be made operational? Assuming efficient markets, an expansionary monetary policy that the public perceived as likely to be effective would immediately raise the prices of actively traded commodities, foreign exchange, and gold. But Keynes himself would argue that the overall price level would respond to monetary shocks with a lag. Thus, under Keynes's own criterion, any monetary policy perceived as being effective would appear as an unfair or competitive devaluation. Conversely, an unintended side effect of his opposition to competitive devaluations would be an opposition to any effective monetary policy.
There is one passage that suggests Keynes's might have had some awareness of this dilemma: "On the American side, we must believe that the President does not really intend an internecine strife of competitive currency devaluation, and does not desire a lower exchange value of the dollar than is justified by the rise in costs, actual and prospective, inside the United States" [Collected Writings, 265-66, italics added]. Under this expanded criterion, Roosevelt could have defended the currency depreciation generated by his gold-buying program on the basis of his intention to restore prices to their 1926 levels. This suggests that Keynes failed to understand the fundamental problems associated with trying to peg real exchange rates.
To summarize, Keynes claimed to advocate expansionary monetary policies during the depression. But he placed so many constraints on these proposals that there was little likelihood that they would be successful in significantly boosting aggregate demand. Keynes's pessimism on monetary policy may have resulted from a combination of his understanding of the limited results achievable under his proposals and his misunderstanding of what might be achieved by (fiat money) policies that he never seriously considered. The irony is that in his theoretical work Keynes became an unwitting prisoner of the gold standard mindset against which he had fought so vigorously.
VII. MONETARY REGIMES AND KEYNESIAN THEORY
Viewing the General Theory as a gold standard model can account for other key aspects of the Keynesian revolution. To do so we first need to examine what parts of this revolution were linked to the monetary ineffectiveness proposition and the ways in which those aspects of Keynesian macroeconomics have changed in response to the gradual elimination of gold from the international monetary system. If many of the most distinctive features of 1930s Keynesianism were in fact a product of the gold standard, then the move to fiat money regimes should have had very specific effects on the development of Keynesian theory and policy.
To examine the role of the gold standard in Keynesian theory it will be useful to first imagine a monetary regime that Keynes himself did not advocate but that would be viewed sympathetically by modern New Keynesians. Consider a pure fiat regime where the central bank continuously adopted a policy of setting the money supply (or interest rates) at a level that was expected to generate sufficient nominal spending to assure full employment with only modest inflation. For instance, the central bank might adopt a 5% per year nominal GDP growth target.(7)
Now consider some of the most distinctive features of the General Theory. Referring to the famous "paradox of thrift" and its implications for bourgeois values(8) Schumpeter [1970 (1951), 290] noted "This is what the Keynesian Revolution amounts to." Note that the paradox of thrift is easily reconcilable with a gold standard regime. Under a gold standard regime, increased saving can undoubtedly reduce velocity (through lower interest rates) and hence reduce nominal spending. But how could an excessive propensity to save produce secular stagnation under a fiat money regime where the central bank is targeting the growth rate of nominal spending?
The General Theory is also notable for its unorthodox policy recommendations, particularly its partial rehabilitation of the theory of mercantilism. Throughout most of his career, Keynes was a free trader. But just prior to the 1931 devaluation, Keynes saw the international gold standard as a constraint that prevented the Bank of England from adopting an expansionary monetary policy to promote recovery and recommended the imposition of a tariff as a second-best policy for limiting gold outflows and thus permitting monetary expansion. After Britain left gold in September 1931, this recommendation was dropped.
Although many economists advocated expansionary fiscal policies during the 1930s, Keynes took the extreme position that deficit spending was necessary because monetary policy was incapable of restoring full employment.(9) As with the paradox of thrift and protectionism, it is difficult to see how fiscal policy could play the role envisioned by Keynes in an economy where the central bank was targeting expected nominal spending. Perhaps if there is a shorter time lag for fiscal policy it could play a useful supplemental role. But Keynes's views were not based on differential lags, and in any case most economists would accept Laidler's [1990, 99] assertion that lags are now seen as a greater problem for (discretionary) fiscal policy: "The tax-cuts implemented in the United States in 1964 were at the time hailed as a triumph of Keynesian policy, but they were proposed in 1962. To hold up a measure that took two years to implement as an example of active stabilization policy is preposterous, for what this episode brings home is that, in the United States, such policy has to be carried out by monetary means or not at all. . . . faith in the efficacy of monetary stabilization policy has come to be regarded as the hallmark of 'Keynesianism', and a denial of that efficacy a characteristic 'anti-Keynesian' position."
Patinkin [1982, 203-9] noted that in 1929 Keynes had advocated an expansionary fiscal policy for Britain, but a only year later in the Treatise had instead advocated monetary stabilization policies. He suggested that in the earlier book Keynes was addressing a domestic audience and treated the constraints imposed by the international gold standard as exogenous. Conversely, he had emphasized monetary policy in the Treatise because he was seeking a larger (global) audience and hoped to influence the operation of the international gold standard.
Even the preceding explanation, however, cannot fully account for Keynes's preference for fiscal policy. In the Means to Prosperity, Keynes argued that only an expansionary fiscal policy could raise world prices but also admitted that this policy might conflict with the need for lower interest rates: "It is at this stage that a certain dilemma exists; since it may be true, for psychological reasons, that a temporary reduction of loan-expenditure plays a necessary part in effecting the transition to a lower long-term rate of interest. Since, however, the whole object of the policy is to promote loan-expenditure, we must obviously be careful not to continue its temporary curtailment a day longer than we need" [1982 (1933), Collected Writings, vol. 9, 35354]. This is an extraordinary admission and suggests that a basic incoherence remains in Keynes's attitude toward fiscal and monetary solutions to depression, even if one accounts for the constraints imposed by the international gold standard.
VIII. FROM THE GOLD STANDARD TO PURE FIAT MONEY REGIMES
One would expect any economist's views to be influenced by the policy regimes with which he or she was most familiar. If the preceding analysis has any value, it is not as a critique of Keynes but rather as a framework for understanding the subsequent development of Keynesian macroeconomics as well as the post-General Theory exegesis. In this section, I offer some observations on how the gradual removal of gold from the world monetary arena has impacted Keynesian economics.
Bretton Woods complicates any attempt to draw a sharp distinction between commodity and fiat money regimes and also complicates the analysis of the impact of regime change on Keynesian economics. Because the dollar price of gold remained fixed between 1934 and the late 1960s, there is a sense in which the United States (and for part of this period, much of Europe) remained on the gold standard. But it is also true that institutional changes allowed the world's stock of monetary gold to support a much larger money stock; hence, the price level (in gold terms) increased substantially during this period. By the late 1960s, the gains achievable from economizing gold reserves were probably nearing an end, and central banks were faced with the choice of either submitting to the sort of constraints associated with the classical gold standard, or allowing an increase in the price of gold.
Although Bretton Woods was not equivalent to the classical (or even interwar) gold standard, it was similar enough that for much of this period Keynesian economists continued to place little emphasis on inflationary expectations. By the late 1960s, however, inflationary expectations had risen to the point where macroeconomists began to pay more attention to issues such as the Fisher effect and the instability of the Phillips Curve relationship.
Leijonhufvud [1968, 19-20] noted that "Keynes" judgments of what was politically feasible at the time has colored his writings to an extent that more academically oriented economists have not always fully appreciated." He did not, however, argue for the central role of the international gold standard regime as the critical factor in the formation of Keynesian macroeconomics. Given the fact that Leijonhufvud's analysis coincided with the end of a 89-year period during which the dollar price of gold was adjusted only once, it is hardly surprising that he would not single out a fixed gold parity as the notable feature of Keynes's world.(10) Bretton Woods was still a quasi-gold standard, and only with the demise of that regime are we able to get a clear view of the importance of gold to Keynes and his contemporaries.
Six years later, Bretton Woods had collapsed and in a discussion of Keynes's views on the limits of monetary policy under a gold standard, Moggridge and Howson [1974, 235, footnote] noted: "To modern eyes the absence of the possibility of exchange rate adjustment in such a case seems odd. However, in the 1920s overnight exchange parity changes of the type with which we are familiar were unknown among the major countries in the international economy." Conversely, by 1988 inflation was so deeply ingrained in the psychology of the public (and economists) that Meltzer could suggest using inflation as way out of the liquidity trap. What is so interesting about Meltzer's observation is that it highlights how little recognition that there has been as to the almost total inapplicability of the key elements of 1930s Keynesianism to a world of fiat money regimes. Twenty-seven years after his seminal work on Keynes, Leijonhufvud [1995, 1341] noted, "The obvious instability of the Phillips relation in the 1970s tilted the balance of professional opinion sharply against Keynesian economics and in favor of monetarism. (It is more seldom noted that its apparent stability before that time might cut the other way.)" In natural rate models the Phillips Curve remains stable unless there is a change in expected inflation. Not surprisingly, the Keynesian model thrived under a monetary regime that generated near zero inflation expectations, and was eclipsed when most countries shifted to more inflationary fiat money regimes.
Where does this leave Keynesian economics in the 1990s? Certainly the importance that Keynes gave to the "underworld" of economics (that is, protectionism, the paradox of thrift, and so forth), has largely been rejected, at least by the New Keynesians.(11) And ever since the breakdown of Bretton Woods, there has been a steady increase in the perception that central banks have both the ability and the responsibility to keep nominal GDP growing at a satisfactory rate. Not surprisingly, this perception (as well as the problem of policy lags) has resulted in fiscal policy being relegated to a subsidiary role in stabilization policy. In many countries, discretionary fiscal policy issues are now framed in terms of their impact on long-run saving and investment trends (that is, the classical approach that Keynes opposed)u
Given the preceding analysis, it may seem surprising that Keynesian economics continues to thrive in the 1990s. In fact, one of the most important innovations of the General Theory, the replacement of the monetary approach to aggregate demand with the expenditure approach, seems to have survived the demise of the gold standard. But the advent of the expenditure approach coincided with the shift of emphasis to fiscal policy and with the development of the paradox of thrift, concepts that are easier to visualize using that approach. If the profession continues to deemphasize these aspects of Keynesian economics and continues to focus more and more on monetary stabilization policy, then a revival of the monetary approach to aggregate demand is not inconceivable. Were that to occur, there would be essentially nothing left of the Keynesian revolution. Yet Keynesian economics would continue to thrive as members of the (New) Keynesian School continued to redefine themselves as proponents of sticky-price models with a self-correcting mechanism. The irony of this, of course, is that these are precisely the sort of models that Keynes was rebelling against during the 1930s.
In some respects, the thesis presented in this paper runs directly counter to the standard interpretation of the General Theory. Patinkin [1982, xxiii] concisely stated the mainstream view of the role of the gold standard in the development of Keynes's work: "Similarly, the predominant emphasis of the Treatise on Money (1930) on the problems of unemployment and of the workings of the international gold standard reflected the major economic concerns of the period, particularly in Britain. . . .[In contrast] the General Theory  is concerned almost exclusively with the problem of mass, long-run unemployment in a closed economy - that is, one not subject to the restrictions imposed by the gold standard." It may be useful, therefore, to review why the General Theory is best thought of as containing a gold standard model, rather than a "fixed-price" or "depression" model of a closed economy.(12)
The perception that Keynes developed a "depression model" seems to be linked to the view that Keynes's policy prescriptions were intended for an economy operating at below full employment. There is no evidence, however, that under a fiat money regime, a suitably expansionary monetary policy would be incapable of expanding the nominal output of an economy in a deep depression. And whereas there is little theoretical (or empirical) support for the view that monetary policy is ineffective during depressions, there is abundant theoretical support for the view that monetary policy may be ineffective in an economy tied to an international gold standard.
I have emphasized the role of stable price expectations in the General Theory. An obvious question, then, is whether there is any advantage to seeing the General Theory as a gold standard model, rather than simply as a fixed-price model. As noted earlier, under the pre-World War II international gold standard the expected rate of inflation was probably near zero. Yet the actual price level was highly procyclical, particularly during 1914-36, and thus one would hardly expect Keynes to have developed a fixed-price model. This pattern is mirrored in the General Theory, Keynes clearly incorporates inflation into his model (in the special case of full employment), but almost never wavers from the presumption that the expected rate of inflation is zero (that is, the pattern observed under a gold standard).
If the preceding distinction represented the only flaw in the traditional view then we could simply replace "fixed-price model" with "fixed-expected-price level model." But there is more reason to focus on the role of the gold standard than simply its implication that inflation expectations can be safely ignored. In the General Theory, Keynes's discussions of monetary policy ineffectiveness made several references to the situation in the U.S. during the early 1930s. In hindsight, it is clear that he was not referring to a "liquidity trap" as we currently use the term but rather to a situation where constraints imposed by the international gold standard might prevent expansionary policy steps from increasing the money supply.
During the period between the Treatise and the General Theory Keynes did shift his focus toward a closed-economy setting. But this is best viewed as shifting from a perspective where there was still hope for international policy coordination under a gold standard regime, to the perspective of a economy operating under a quasi-gold standard with little hope for policy coordination. Admittedly, Keynes did not consciously view the General Theory as a gold standard model. But he continued to assume, and to advocate, policies that would, in effect, assure reasonable stability in the price of gold. And his increasing pessimism about the prospects for effective monetary policy was based on his observations of real-world policy failures that were linked to the operation of such a system.
Some of Keynes's contemporaries understood that the gold standard influenced his writings. For instance, Patinkin [1981, 301] quotes Henry Simons (in a 1936 review of the General Theory) as observing that "Indeed, if the whole book could be interpreted simply as a critical appraisal of the traditional gold standard, implemented through central-bank operations, one's judgment of its main ideas might be extremely favorable." It is interesting that Simons sees Keynes's work as a critique of the traditional gold standard. From a modern perspective, Keynes often seems more a prisoner, rather than a critic, of the gold standard framework.
Ultimately, Keynes's pessimism regarding monetary policy effectiveness was a failure of imagination. Unlike Irving Fisher, Keynes neither visualized, nor approved of, a highly expansionary fiat money regime to promote reflation. This distinction is not obvious from their theoretical writings, but shows up clearly in their contrasting views of the latter stages of Roosevelt's dollar depreciation program.
Between April 1933 and January 1934, the depreciation of the dollar did raise the U.S. price level sharply, despite the fact the fact that output remained far below capacity and "bottle-necks" were presumably not a serious problem.(13) In the latter stages of the program (under the influence of George Warren), Roosevelt was explicitly following Fisher's recommendation that the dollar be devalued until prices return to their 1926 levels. In the December 1933 letter to the Times cited earlier, Keynes's swipe at the "quantity theory" almost certainly was a reference to Warren and, by implication, Fisher. Whether because of concerns about competitive devaluations, or fears of a pure fiat regime devoid of any nominal anchor, Keynes allied himself with the conservative critics of this highly expansionary policy.(14)
No single approach is capable of fully illuminating a work as complex as the General Theory. Although this paper has emphasized the role of the gold standard, Keynes's focus on the liquidity effect as the sole transmission mechanism for monetary policy undoubtedly contributed to his pessimism on this issue. Other factors not considered in this paper may also have contributed to the broader themes in the General Theory. It is those factors linked to the constraints imposed by the international gold standard, however, that provide 1930s Keynesianism with its most distinctive features.
I thank H. Clark Johnson, Axel Leijonhufvud, David Laidler, Hugh Rockoff, Kevin Stiroh, and several anonymous referees for helpful comments and suggestions.
1. Part of a broader strategy for depreciating the dollar, this program involved having the U.S. government purchase gold at progressively higher prices. McCloskey and Zecher  show that dollar depreciation during 1933 led to rapid increases in the price level.
2. Regarding the policy ineffectiveness proposition, see the General Theory [1964 (1936), 164, 266-67, and 309]. Keynes's views of aggregate supply are clearly described on page 300 of the General Theory.
3. That is, nominal wage rates.
4. Admittedly, the term "devaluation" is somewhat ambiguous. Contemporary writers used this term synonymously with both a depreciation of purchasing power and/or an increase in the official price of gold. In private correspondence, Hugh Rockoff suggested that the term "devaluation" might simply refer to monetary expansion. In that case, Keynes may have intended the monetary ineffectiveness proposition to extend to prices.
5. A $3 billion increase in the monetary base during 1933 would have left nominal income far below 1929 levels, unless base velocity had also increased sharply. Admittedly, velocity often increases during periods of expansion and/or inflation, but it is also probable that the Fed would have partially neutralized the impact of the greenbacks through open market sales of securities.
6. Long-term bonds fell sharply during the first half of November 1933. The U.S. financial press attributed this decline to uncertainty generated by the gold-buying program that was the latest stage of a dollar depreciation policy begun in April 1933. In contrast, bonds rallied on news of Roosevelt's stabilization announcement on 15 January 1934, despite the fact that it implied a small additional depreciation. The financial press attributed the divergent responses to the fact that the latter announcement had placed a definitive limit on the depreciation of the dollar, that is, the step was viewed as a once-and-for-all increase in the price level.
7. As noted in the previous section, there is little evidence that Keynes ever doubted the proposition that, under a pure fiat money regime, an expansionary monetary policy pursued it outrance could generate any degree of inflation, and therefore, nominal GDP growth. Of course, policy lags may complicate such a proposal, but Keynes was primarily concerned with secular stagnation generated by oversaving.
8. Here Schumpeter is referring to values like thrift, and the justification for income inequality - which can contribute to saving.
9. Tavlas  noted that the Chicago School also favored deficit spending during the 1930s, but justified it as a way of generating an expansionary monetary policy. Unlike Keynes, the Chicago School never abandoned the quantity theory mode of analysis.
10. Adjustments of the parity of other key currencies such as the pound were somewhat more frequent, but still more the exception than the rule.
11. This study does not look at the Post-Keynesians. Thies  notes that the fourteenth edition of Samuelson and Nordhaus's famous text contains no mention of the paradox of thrift, one of the key concepts in the General Theory.
12. Many textbooks regard the Keynesian model as a fixed-price model. Meltzer [1988, 202] quotes Richard Kahn as concluding that the General Theory was "designed for an economy in a state of depression."
13. Consumer prices rose about 10% and wholesale prices about 20%.
14. Note, however, that Keynes could have responded to competitive devaluations in the United States by recommending that Britain do the same. In the Keynesian model, expansionary monetary policies are not a zero-sum game.
Barsky, Robert B. "The Fisher Hypothesis and the Forecastability and Persistence of Inflation." Journal of Monetary Economics, January 1987, 3-24.
Bernanke, Ben S. "The Macroeconomics of the Great Depression: A Comparative Approach." Journal of Money, Credit and Banking, February 1995, 1-28.
Keynes, John M. A Tract on Monetary Reform. London: Macmillan, 1923.
-----. A Treatise on Money. London: Macmillan, 1930 .
-----. The General Theory of Employment, Interest, and Money. New York: Harcourt, Brace & World, 1936 .
-----. The Collected Writings of John Maynard Keynes. Volume 9, edited by Donald Moggridge, Cambridge: Cambridge University Press, 1982.
-----. The Collected Writings of John Maynard Keynes. Volume 13, edited by Donald Moggridge, Cambridge: Cambridge University Press, 1982.
-----. The Collected Writings of John Maynard Keynes. Volume 21, edited by Donald Moggridge, Cambridge: Cambridge University Press, 1982.
Laidler, David. Taking Money Seriously and Other Essays. Cambridge: MIT Press, 1990.
Leijonhufvud, Axel. On Keynesian Economics and the Economics of Keynes. New York: Oxford University Press, 1968.
-----. Review of Money and the Economy' Issues in Monetary Analysis by Karl Brunner and Allan Meltzer, in Journal of Economic Literature, September 1995, 1341-43.
McCloskey, Donald N., and Zecher, J. Richard. "The Success of Purchasing Power Parity: Historical Evidence and Its Implications for Macroeconomics," in A Retrospective on the Classical Gold Standard, 18211931, edited by Michael B. Bordo and Anna J. Schwartz. Chicago: University of Chicago Press, 1984, 121-70.
Meltzer, Allan. Keynes's Monetary Theory: A Different Interpretation. Cambridge: Cambridge University Press, 1988.
Moggridge, Donald E., and Susan Howson. "Keynes on Monetary Policy, 1910-1946." Oxford Economic Papers, July 1974, 226-47.
New York Times. "From Keynes to Roosevelt: Our Recovery Plan Assayed," 31 December 1933, Sec. XX, p. 6.
Patinkin, Don., Essays on the Chicago Tradition. Durham, N.C.: Duke University Press, 1981.
-----. Anticipations of the General Theory and Other Essays on Keynes. Chicago: University of Chicago Press, 1982.
Schumpeter, Joseph A., Ten Great Economists From Marx to Keynes. New York: Oxford University Press, 1951 .
Tavlas, George S. "Keynesian and Monetarist Theories of the Monetary Transmission Process." Journal of Monetary Economics, May 1981, 317-37.
Thies, Clifford F. "The Paradox of Thrift: RIP." Cato Journal, Spring/Summer 1996, 119-28.
|Printer friendly Cite/link Email Feedback|
|Date:||Jul 1, 1999|
|Previous Article:||Are 'risky assets' substitutes for 'monetary assets?'.|
|Next Article:||Rent-seeking firms, consumer groups, and the social costs of monopoly.|